A surface analysis by laser ionization (SALI) apparatus has been used to obtain, for the first time, real-time photoionization mass spectra of the shear-induced molecular-fragment emission from an explosive. NTO (5-nitro-1,2,4-triazol-3-one) was chosen for these experiments because of its potential utility as a reasonably energetic, but very insensitive, explosive. Using vacuum ultraviolet single-photon ionization, the shear-induced NTO spectra were obtained with a spring-driven shearing device installed in a SALI chamber directly beneath the mass spectrometer sampling region. For comparison, we also obtained spectra under either slow-heating or rapid pulsed-laser heating conditions. The shear-induced spectra are dominated by a peak at m/z 99, which is not seen in the thermal- or laser desorption spectra. This peak is assigned to the closed-shell traiza-diketone produced by a nitro-nitrite rearrangement, followed by NO loss and then by rapid bimolecular H-atom removal. The stability of the cyclic diketone intermediate thus generated could help to explain the shock insensitivity of NTO. Laser-desorption spectra were also obtained both on fresh NTO samples and on samples that have been recovered from marginally sub-critical drop-weight impact tests. Comparison of spectra obtained with and without laser desorption, and as a function of temperature, demonstrate that the sequences of fragment ions observed under laser desorption conditions are the result of thermal decomposition, not of ion-fragmentation. The sequence of thermally generated fragments is dominated by M-16, M-30, M-45, M-46, and M-59. This series suggests several decomposition pathways, dominated by the same nitro-nitrate rearrangement and NO loss as the shear-induced decomposition. However, under the lower-density, but higher temperature, thermal or laser-desorption conditions, subsequent bimolecular H-atom removal to produce the closed-shell diketone is evidently slower than unimolecular ring-opening adjacent to the carbonyl group. We show how this sequence satisfactorily explains (1) the ＇＇initial＇＇ formation of CO2 that has been previously reported, (2) the results of nitrogen double-labeling experiments, and (3) the fact that neither NO2 nor HONO have been seen as substantial initial products of NTO decomposition.